The platform contrasts pure parametric memory with grounded search results to expose discrepancies between a model's internal training data and live web facts, using the Fact & Gap Checker to classify claims as confirmed, contradicted, or missing knowledge gaps.
flowchart LR
A[Brand Entity] --> B[Parametric Memory Probe]
A --> C[Grounded Memory Research]
B -->|Zero-Shot / Logprobs| D[Recalled Claims]
C -->|Live Search / URL Context| E[Verified Findings]
D --> F[Fact & Gap Checker]
E --> F
F --> G[Confirms / Contradicts / Gaps]